Inside Carlyle's AI rollout: Tech chief shares wins, challenges, and cost savings

Business Insider

16 hours ago

Lucia Soares is helming Carlyle's AI transformation after years of bringing tech to big companies.
She spoke to BI about the firm's AI rollout and how it's already resulting in cost savings.
She also spoke about life as a bicoastal executive and what she learned from her immigrant parents.
Lucia Soares had been working for Carlyle for four years when the private equity giant's CEO called to ask if she would take on a new role.
"I originally focused on using tech to create portfolio value," she told Business Insider, referring to the companies Carlyle controls. "Then, two years ago, our new CEO called me and said, 'Can you please do what you're doing for our portfolio companies but for our own company internally?'
Now, Soares — as Carlyle's chief information officer and head of technology transformation — is taking on a new challenge: Bringing artificial intelligence to the investment giant's 2,300 global employees.
She spoke with Business Insider about the rollout, including the successes, the pitfalls, and how the company is implementing checks and balances. She explained where the company is already seeing cost savings, for example.
She also walked us through her life as a bicoastal tech executive — and how she learned to hustle from a young age, helping her immigrant parents sell plants at the flea market on weekends. The interview has been edited for length and clarity.
In my 27 years in technology, I've learned that you can't start with technology itself as the goal. People said that e-commerce is the goal, or that digital is the goal. Now, they say AI is the goal. And actually it's not.
Instead, we start with our business goals: we want to grow, create efficiencies, and build a strong tech foundation. AI and other technologies are levers to achieve these goals.
Increasing our employees' AI fluency is a strategic priority. They get AI training from the day they start at Carlyle, and are introduced to a wide range of tools they can use.
Now, 90% of our employees use tools like ChatGPT, Perplexity, and Copilot. We also have an AI champions' council where early adopters can play around with tools and eventually share best practices.
We're using AI to transform our workflows through Project Catalyst, which automates processes. We're also developing custom tools that leverage proprietary data to deliver insights instantly—saving investors from sifting through endless materials. Today, Carlyle's credit investors can assess a company in hours using generative AI, instead of spending weeks on research.
How is AI impacting the average worker at Carlyle? Are they required to use the technology?
It depends. Some business leaders have made it a requirement to put all investment committee memos in an AI tool for them to review. Others are not so direct about it, but everybody is seeing how it can make their jobs easier and challenging their teams in meetings to talk about the value they are deriving from AI tools.
As a firm, we have a return-on-investment strategy, and my team aims to deliver a certain amount of ROI every year.
We're not eliminating people's jobs, but we believe that it can help reduce dependency on outside services costs. For example, we can use AI to review legal invoices and catch errors that will reduce our costs. We've seen real savings as a result.
How do you balance autonomy with the risks of adoption?
I think a lot about that. I worry about kids in school using a tool to write an essay and not being able to think. But you have to wonder how people felt when the calculator came out, and if they thought no one would ever be able to do math on their own again.
We never allow AI to make a final decision. There's always a human in the loop, and someone needs to be accountable for the final results.
For example, when employees use AI to write a report, we have employees write a final paragraph summarizing the output to ensure they're thinking critically about it.
Let's start with success.
When investors invest with us, we can at times receive up to 80-page documents with questions about everything from our employees to cybersecurity training. It's very manual.
We had one team decide they'd try to use AI to make investor diligence easier. Despite having just one technologist, this team found a solution to automate the process, which we're launching later this year.
We seek to empower people to solve things themselves, with embedded technologists across the organization.
We experienced more challenges dealing with regulatory restrictions on large language models globally. We learned the hard way that these regulatory hurdles require a lot of evaluation. We're launching solutions, but it's taking longer than expected to deploy.
You might think you can go fast with AI, but it doesn't always work that way, especially in today's global climate.
Has any single piece of career advice stuck with you over the years, and what is it?
Early on, I was advised to always raise my hand for the extra hard assignments. In other words, take a risk and bet on yourself.
My parents are immigrants, and I learned work ethic, courage, and audacity from them. But when I entered the workforce, I had impostor syndrome. With blue-collar parents, the office environment was completely different for me.
By taking on difficult assignments, I created relationships and visibility and was able to learn and grow more.
Tell me about your parents.
They are from the Azores Islands in Portugal. They came to the US during the dictatorship years. My dad only went to school up until the age of 10, because his family could not afford to pay for more education. He can add, subtract, and multiply, but was never taught how to divide.
He came to the US after serving in the Portuguese Army to give his family a better future. He knew no English.
He became a custodian, cleaning schools, and had a side hustle selling house plants at a flea market on the weekends. We all helped cultivate and sell the plants. I learned a lot from my parents.
I am bicoastal: I spend one week a month in DC and also time in New York, but I live on the West Coast and work out of our Menlo Park office.
On the East coast, I might start my day — work permitting — listening to news podcasts, going for a run, meditating, and eating a healthy breakfast.
At home, I start really early in the morning. I don't always get that workout in, but I start with some early calls, and then take a break to drive my daughter to school before heading to the office.
When I get to my desk, I write down the day's priorities. I've done this my whole career, and try not to let constant fire drills overtake those priorities. When you're driving transformation, you have to keep strategy at the forefront.
What are the most important meetings of your week?
The most important meetings are the unplanned ones. For example, I run into a coworker, and we start talking about our kids. Then they bring up a company we should partner with. Or I run into an administrative assistant, and they show me new ways they're using Copilot. I get inspired by solving problems with people in real time.
The second most important meetings are the ones where we drive strategy and brainstorm. As technologists, you can fall into the Dilbert category of employees, where you just work through problem resolutions. So I force strategy onto the calendar to ensure we think big and ambitiously about tech transformation.